About the Project UK river water quality is a pressing political, public and environmental concern: only 15% of English rivers
meet 'good' status under the Water Framework Directive, and water companies and the government face mounting pressure to improve waterways. The picture is similarly stark globally, where freshwater vertebrate populations have declined by 85%, driven by pollution, habitat loss, agricultural intensification and climate change. Freshwater science urgently needs new tools to diagnose declining water quality, attribute it to specific stressors (e.g. sewage pollution, abstraction, diffuse agricultural runoff) and inform effective mitigation. Monitoring is transitioning from 'grab' sampling at weekly to monthly intervals to continuous, sensor-based monitoring at sub-hourly resolution. This shift, exemplified by the UK's Section 82 (Environment Act 2021) programme to deploy 20,000 continuous sensors nationally, generates high-volume, multivariate datasets that reveal short-lived pollution events and natural diurnal cycles invisible to grab sampling. However, existing manual interpretation methods cannot keep pace with this data volume and complexity, creating a critical bottleneck between data collection and actionable insight. This project will develop novel data-scientific methods to interpret multivariate, continuous water quality datasets at national scale. The student will: (1) build data-reduction and dimensionality-reduction approaches that enable large numbers of sensors to be compared systematically, identifying recurring 'typologies' of water quality response; (2) adapt event-detection methods from bioacoustics and seismology to automatically identify discrete events, such as sewage spills or storm-driven pollution pulses, within continuous sensor records using machine learning; and (3) apply these methods to a UK-wide compilation of sensor datasets, assembled by the student, to characterise the land-use, climatic and hydrological drivers of water quality nationally for the first time. The project directly addresses NERC's priorities in environmental data science and freshwater health, and translating research into environmental management practice, responding to an urgent need created by the UK's rapid expansion of continuous monitoring infrastructure. Working closely with CASE partner The Rivers Trust, the student will co-develop a practitioner toolkit that translates these methods into a standardised workflow for extracting insight from continuous monitoring data. Needs identified by the CASE supervisor will steer project development, ensuring the research responds directly to practitioner requirements. The resulting toolkit will be deployed across The Rivers Trust's network of catchment managers, providing a direct route from methodological innovation to real-world impact. During their 6 month placement with The Rivers Trust, the student will bring their expertise in continuous monitoring data science to the organisation's data science team, contributing to an active project with Defra and the water industry on establishing 'baseline' water quality conditions against which pollution events can be identified, a key question for optimising the future deployment of continuous sensors nationally. Specialist expertise in continuous monitoring and data science is essential to this work, and the student will be uniquely placed to fill this gap, gaining direct experience of how research translates into environmental policy and practice. Supervisors: Alex Lipp (UCL) Matthew Fox (UCL) Andrew Johnson (Rivers Trust) Funding Notes: Fully funded place including home (UK) tuition fees and a tax-free stipend at UKRI rates. Additional funding to cover full overseas fees is available for a maximum of 6 studentships. Enquiries email name and address: TREES.Admissions@ucl.ac.uk Application Web Page: https://www.trees-dla.ac.uk/apply


